FPCO-Dialog: A Multi-Turn False-Premise Benchmark for Correction and Cooperation in Vision-Language Models
FPCO-Dialog is introduced, a benchmark for evaluating correction and cooperation behavior in VLMs under repeated false premises, and reveals substantial and persistent cross-model differences in aggregate correction tendency, model-specific turn-wise dynamics, and systematic variation across false-premise types under t...
Jiayuan Ma, Yu-Qi Lu, Wei-Yang Guo et al.
· 0 citations